> MBAs are just trained to optimize around the the current maximum in the "energy" landscape. These maxima are all too often not the best in the landscape, or even worse, can vanish over time as the landscape changes.

> MBAs are like a gradient optimization algorithm without any simulated annealing process.

Now that is content-free babble. If you are a bot, more work is needed. If you are a human, you should be ashamed of yourself.

He attempted to explain why MBA behaviour is rational using the language of hill climbing algorithms: http://en.wikipedia.org/wiki/Hill_climbing

He didn't express any big new insights, but there was an honest attempt to impart information behind that post.